The submitted manuscrlpt has been authored by a contractor o f the U S Government under contract No W 31.109-ENG-38 Accordingly. the U S Government retalns a nonexclusive. royalty-tree license to publlsh or reproduce Ihe published form of thlS contribution. or allow others to do so. for U s Government purposes

UTILIZATION OF OXYGEN-ENRICHED AIR IN DIESEL ENGINES: FUNDAMENTAL CONSIDERATIONS eotu F - + T O 4 1

Debesh Lahiri and Pramod S. Mehta Indian Institute of Technology

Madras, India

Ramesh B. Poola and Raj Sekar Argonne National Laboratory

Argonne, Illinois

ABSTRACT Utilization of oxygen-enriched air in diesel engines holds

potential for low exhaust smoke and particulate emissions. The majority of the oxygensnriched-air combustion-related studies so far are experimental in nature, where the observed results are understood on an overall basis. This paper deals with the fundamental considerations associated with the oxygen-enriched air-fuel combustion process to enhance understanding of the concept. The increase in adiabatic flame temperature, the composition of exhaust gases at equilibrium, and also the changes in thermodynamic and transport properties due to oxygen-enrichment of standard intake air are computed. The effects of oxygen-enrichment on fuel evaporation rate, ignition delay, and premixed burnt hc t ion are also evaluated. Appropriate changes in the ignition delay correlation to reflect the effects of oxygen- enrichment are proposed. The notion of oxygen-enrichment of standard intake air as being akin to leaning of the fuel-air mixture is refuted on the basis of the findamentally different requirements for the oxygen-enriched combustion process.

INTRODUCTION Current engine design and development strategies revolve

around the need to reduce engine emissions to legislated limits. Engine-out emissions serve as a pointer to the efficacy of in- cylinder combustion-related processes. The concept of oxygen- enrichment aims at limited substitution of the nitrogen in air by oxygen to achieve low emission levels. Because of the increased oxygen content, additional fuel is burned. The resulting increase in power output is a beneficial offshoot, though it is not attempted for its own sake. Oxygen-enrichment of combustion air provides an opportunity to achieve ignition with minimum amounts of premixed fuel. because it reduces the ignition delay

RECEI

O S T I

period under all operating conditions. As a result, the peak cylinder pressure is low. Oxygen-enrichment of standard intake air also promotes combustion with alternative fuels, as well as low-grade and water-emulsified fuels (Sekar et al. 1990). However, oxides of nitrogen (NO,) emissions would be higher with oxygen-enrichment because of the accompanying higher combustion temperatures. The level of technological complexity, economics, and NOx emissions limit govern the oxygen- enrichment level. Complete substitution of air by oxygen is not attempted, despite its theoretical superiority, because of the economics of oxygen generation, technical difficulties associated with implementation in internal combustion (IC) engines, and metallurgical considerations.

The concept of oxygen-enrichment has its genesis in spark ignition (SI) engine applications (e.g., Wartinbee 1971; Quader 1978). At present, attention is more critically focused on compression ignition (CI) engines because of the diverse fields of application. Studies concerning effects of oxygen-enrichment on both Direct-Injection (DI) and Indirect-Injection (IDI) diesel engines have been carried out, with particular emphasis on the DI variety. Ghojel et al. (1983) and Reader et ai. (1995) have studied the effects of oxygen-enriched combustion air in an ID1 diesel engine. Ghojel et al. (1983) varied the oxygen content in the intake between 21 and 40 percent by volume and evaluated the engine performance and emissions at constant engine speed and injection timing while changing the fueling rate. They reported a large decrease in ignition delay period, reduced combustion noise, decreased HC and CO emissions, and a substantial drop in smoke levels. The NOx emissions were found to increase proportionately with oxygen-enrichment. They observed that fuel consumption, power output, and exhaust temperature remained constant, which was rather unexpected in the light of the usual effects of oxygen-enrichment. The findings of Reader et al. (1995) were in complete agreement with those Of

DISCLAIMER

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or use- fulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any spe- cific commercial product, proms, or service by trade name, trademark, manufac- turer, or otherwise does not necessarily constitute or imply its endorsement, m m - mend;rtion. or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not neccssarily state or refiect those of the United States Government or any agency thereof.

.

Ghojel et ai. (1983). A significant contribution of Reader et al. (1995) is the attempt to correlate NOx emissions with the ignition delay period and combustion-generated noise with an equivalence ratio based on oxygen.

The use of oxygen-enriched air is considered as a potential method for soot emission control in DI diesel engines. The reduction in ignition delay period with oxygen-enrichment provides an opportunity to control NOx emissions by retarding the injection timing. Also, the superior combustion characteristics of oxygen-enriched air allow for water injection, which would lower the in-cylinder gas temperatures, thus controlling NOx A review of published literature (Karim and Ward 1968; Iida and Sat0 1986; Sekar et al. 1990; Watson et al. 1990; Virk et al. 1993) reveals that the experimental studies have focused on evaluating the engine performance, combustion, and emissions aspects for varying levels o f oxygen-enrichment, different grades of fuel, and water-emulsified fuels. Unanimity is found in the experimentally observed trends reported by different investigators. However, very limited fundamental bases have been provided to explain these observations. Hence, a need to understand the fundamental aspects of the oxygen-enrichment concept as applied to diesel engines has motivated the present work.

BASIS OF PRESENT WORK In engine-related combustion, the induced air is presumed to

be “dry ambient air.” The following definitions are commonly employed for indicating the quality of the fuelair mixture (Heywood 1988):

Relative air-fuel ratio (A) =

Equivalence ratio (4) = li-1 =

From these parameters, the mixture quality is decided as follows:

(1)

(2)

actual air-fuel ratio stoichiometric air-fuel ratio

actual fuel-air ratio stoichiometric fuel-air ratio

4 1 and li > 1 : fuel-lean mixtures 4 = n = 1 : stoichiometric mixture 6 > 1 and li < 1 : fuel-rich mixtures

Fundamental Considerations The fundamental concept of oxygen-enriched combustion lies

in altering the composition of standard ambient air by increasing the relative percentage of oxygen at the expense of nitrogen. In order to embrace a broader understanding of charge modification with air composition, which includes processes like dilution of intake charge with inert gases, water addition, and exhaust gas recirculation, it would be appropriate to use the term “reformed air” (Maxwell et al. 1993) to cover these effects. The air quality can best be described by the ratio of mole fraction of nitrogen to oxygen. For a given engine displacement and amount of intake air delivered, there exist three different operating regimes, based on the contents of oxygen and fuel in the cylinder, as given below: 1. For a constant fuel input, increasing the oxygen content in

the air by proportionately decreasing nitrogen concentration in air such that

Increasing both the oxygen content in the air and the fuel input while proportionately decreasing nitrogen concentration in the air such that

Increasing the oxygen content of air by proportionately decreasing the nitrogen content in the air and disproportionately increasing the fuel input such that

(The subscripts, amb and OE, stand for conditions with ambient and oxygen-enriched combustion air, respectively.)

Varying the composition of air in terms of oxygen and nitrogen contents is expected to cause changes in its thermodynamic and transport properties, ignition delay, flame temperature, exhaust gas composition, and fuel-air mixing characteristics that may be o f importance in furthering the understanding of the oxygen-enrichment concept. These characteristics, evaluated by using appropriate correlations, are described in the subsequent sections.

Tbermodvnamie Properties Density: In the case of oxygen-enrichment, the definition of

volumetric efficiency is modified in the presence of additional oxygen content in the air. This can be described as apparent volumetric efficiency (AVE) and is defined as:

Volumetric flow rate of oxygen-enriched intake air Rate of change of engine cylinder volume

AVE =

(6) The AVE as defined in Equation 6 includes the effect of manifold temperature on charge density. Both the altered compositions of intake air and higher in-cylinder temperatures due to oxygen- enrichment lead to changes in the density of the charge. The change in density has so far been attributed to higher in-cylinder temperatures due to oxygen enrichment, which is applicable in the quasi-steady working period of the engine cycle. However, at a given temperature prior to combustion, the relative molecular mass (RMM) of air changes with the oxygen-enrichment, and thus, the density. The variation in density with various levels of oxygen-enrichment was computed on the basis of RMM.

Specific Heat Capacity: The effect of reformed air composition on the ratio of specific heats is not reported in the literature. A preliminary calculation is needed to qualitatively estimate the change, following Kothandaraman and Subramanyam (1989).

Transport Properties The two transport properties of importance are viscosity and

thermal conductivity, which influence the flow and heat-transfer- related processes. Due to oxygen-enrichment, the overall temperature after combustion will be raised as a consequence of

2

higher flame temperature. This would increase the heat transfer losses and cooling requirements in the engine. The estimation of transport properties was attempted to highlight the significance of oxygen-enriched combustion. The dynamic and kinematic viscosities, as well as the thermal conductivity, were computed by following Kothandaraman and Subramanyam (1989). The mixture’s thermodynamic and transport properties were estimated. using the square-root rule, by following Spalding (1 963).

Flame Temperature and Equilibrium Composition In the present work, a computer program developed by Olikara

and Borman (1975) was used, with suitable modifications necessitated by oxygen-enrichment, to compute adiabatic flame temperature and equilibrium composition. The change in molar ratio of nitrogen-to-oxygen is included with oxygen-enrichment. The scheme limits the number of product species to 12. However, it gives a fair degtee of accuracy without obscuring the physics of the problem. The adiabatic flame temperature and equilibrium composition were calculated with different equivalence ratios and at various oxygen-enrichment levels. The difference between oxygen-enrichment and lean bum combustion is explained through equilibrium composition calculations.

Phase Change (Fuel Evaporation) Rate The transfer number concept, suggested by Spalding ( 1963), is

used as the basis to compute the phase change rate. The transfer number (B) is defined as the difference in the values of any conserved property in terms of species concentrations in the gas stream and the adjoining stream. This transfer number, which is a driving force for mass transfer, can be understood as being analogous to the potential difference required for the flow of electric current. When the transfer number is estimated by using mass conservation equations, it is called the mass transfer number, Bm; if estimated by using enthalpy conservation equations, it is called the enthalpy transfer number, Bh. These transfer numbers can be estimated with and without combustion. A transfer number greater than zero implies mass transfer into the considered phase (i.e., evaporation of fuel and transfer of fuel vapors into air), whereas a transfer number less than zero implies condensation of air and dilution of fuel, a condition not realized in IC engine combustion.

The evaporation and envelope flame (diffusion) combustion of a fuel droplet are frequently treated by assuming a slowly varying sequence of steady states (i.e., quasi-steady assumption). However, in an engine combustion chamber, the conditions at the gadliquid interface may undergo appreciable fractional changes in a period of the order of the characteristic difision time R,2/Dfu, where R, is the droplet’s initial radius at that instant and Dfu is the diffusion coefficient of the fuel. For example, a 100-pm diesel fuel droplet requires a characteristic diffusion time of a millisecond or so. Even when the vapor composition and temperature are essentially constant at the droplet surface, the starting transient may occupy an appreciable fraction of the total droplet lifetime. These transient effects become quite important when local conditions at the vapor/liquid interface approach thermodynamic critical conditions. The transfer numbers for droplet evaporation with and without combustion are estimated

using a transient analysis. The fuel’s partial pressure is calculated following Rosner and Chang (1973), and the one-third rule of Sparrow and Gregg (1958) is used to calculate the properties at reference temperature in transfer number calculations (Mehta et al. 1985).

&nition Delay Prediction of the ignition delay period for different engine

geometries under various operating conditions and fuel types has been investigated for a long time. However, the variation of air composition on ignition delay has not been a common feature of most of the earlier correlations. From the many existing correlations, only those formulated by Fujimoto (1980) and Hiroyasu (1985) have explicitly considered the effect of oxygen concentration in the intake air in terms of the partial pressure of oxygen (Po2), as given below:

(7)

where a = 0.0276 and Ea/R = 7280 K.

Diesel fuel (C,,HZ6) with a Cetane number of 45 was employed in deriving the above correlation. When used in IC engines, this correlation requires an input of cylinder temperature and pressure. These values could be suitably taken at the end of compression. However, a better estimate is possible by incorporating the injection timing effects, which can be reflected through the values of temperature and pressure at the instant of injection. Also, from the operating regimes described in Equations 3 through 5, the contents of oxygen and fuel in the cylinder are key components when oxygen-enriched combustion air is used. Hence, they should be considered for a realistic estimate of ignition delay period. In the present work, the ignition delay correlation suggested by Hiroyasu (1985) was modified to provide a consistent understanding of the oxygen- enrichment and to reflect the changes in fuel-to-oxygen ratio due to oxygen-enriched air as:

where FOR is the local fuel-to-oxygen ratio (on a mass basis), nl is a constant, p and T are taken at the instant of fuel injection, and the values of ’a’ and Ea/R are the same as in Equation 7. The value of nl is determined by curve-fitting the experimental data. On the basis of experimental data obtained on a single-cylinder Caterpillar engine (Sekar et al. 1991), an average value of -1.15 is recommended for n I .

Fuel-Air Mixing Rate - Multizone Approach The fuel-air mixing process in diesel engines is significantly

slower than the kinetics of fuel oxidation during the greater portion of the combustion process; therefore, the dominant combustion mechanism in the diesel engine is mixing-controlled. The heterogeneity of the fuel-air mixture before the onset of ignition causes the local fuel-air ratio to be extremely low in the regions where fuel evaporation occurs. Initial ignition has the advantage of high oxygen concentration around molecular groups

3

of fuel. To anticipate the pattern of heterogeneity in the charge is an important prerequisite for the development of an efficient combustion system (Andree and Pachernegg 1969). The present work pursues this approach further to predict the temporal and spatial distribution of the fuel-air mixture. DI diesel engines employ multi-hole injector nozzles and high injection pressures, thereby attaining a higher fuel mass flow rate and a larger fraction of fuel in the spray fringes, where possibilities of premixing are higher. The ignition delay time is generally longer than the mixing time, so ignition delay controls the premixed phase combustion. The premixed combustion fraction is evaluated on a more consistent basis by considering the amount of fuel injection and the mass fraction of ignitable mixture prepared during the ignition delay period only. At any instant, these mixture fractions can be put in one of four categories to relate them to engine combustion and emission characteristics:

Overmixed zone Ignitable fuel-lean zone Ignitable fuel-rich zone Undermixed zone

: (4 5 0.3) : (0.3 Q < 1.0) : (1.0 < 4 < 3.0) : (0 > 3.0)

At the end of the ignition delay period, when the in-cylinder pressure and temperature attain values that can sustain combus- tion, the mass fraction of ignitable fuel prepared during the ignition delay period is consumed. The unignitable mass fraction that is outside the flammability limits has to undergo further mixing with the surrounding air and the products of combustion contributed from the burned mass fraction. Since the mixing quality can be identified in terms of ignitable (rich and lean) and unignitable mass fractions at any instant during the spray mixing and combustion process, they are estimated here to determine the effect of oxygen-enrichment on the mixture preparation prior to ignition. The flammability limits for diesel combustion are generally taken as 0.3 < 0 < 3.0 (Shahed et al. 1975) through an approximation with oxygen-enrichment. The oxygen flammability limits are known to be wider than the air flammability limits. The lean limit for flammability is unaffected at higher pressures and with oxygen-enrichment, but the rich limit increases in these cases. The values of rich limits with oxygen-enrichment are not readily available for diesel fuels; hence, the values associated with standard air flammability limits continue to be used.

In order to reflect the increased complexities associated with oxygen-enrichment, the local fuel-air mixture quality is predicted by a time-marching process model. A multizone, phenomenological model is formulated, using appropriate process correlations. Spray growth is computed by using the spray penetration equation suggested by Dent (1971). A semi- cone angle is taken as 1Io, as recommended by Abramovich (1963) for a non-burning fuel spray. Fuel-air mixing calculations are based on mean injection velocity and an air entrainment rate equation, following Ricou and Spalding (1961). The multizone configuration of the spray is arrived at by dividing the fuel injection process into a number of intervals of lo CA duration and assuming the distribution of each of these packets follows a similarity law for concentration of fuel as proposed by Abramovich (1 963). The details of the complete model can be found in Mehta et al. (1996). This modeling procedure provides

a fuel-air distribution in terms of the local equivalence ratio in each zone. With an assumption of uniform cylinder pressure and temperature prior to combustion, these equivalence ratio distribution histories form the basis to estimate the amount of mixture prepared for burning under stoichiometric conditions. Depending upon the local equivalence ratio and cylinder pressure and temperature conditions, the local burning temperature, expressed as adiabatic flame temperature in various spray zones, is computed following Olikara and Borman (1975). The droplet size distribution in terms of Sauter mean diameter is included. The detailed droplet evaporation calculations are avoided by using the assumption of instantaneous evaporation. The model does not include the effects of swirl on fuel-air mixture formation. The ignition delay is either an input value or can be estimated by using Hiroyasu's correlation as expressed in Equation 7. The engine details and other inputs used in the calculations are specified in Table 1.

Table 1. Specifications of the test engine

Bore (mm) 103.29 Stroke (mm) 120.65 Compression ratio 16 : 1 Engine speed (rpm) 1600 Injection conditions Mean injection pressure (bar) 250

Number of injector holes 4 Start of injection (deg. btdc)

Orifice diameter (mm) 0.23

10 25 78.2

Duration of injection (deg. CA) Fueling rate per cycle (mm3)

Characterization of Premixed Combustion Phase Diesel combustion is essentially a mixing control process.

However, the role of the premixed combustion phase in diesel engines has acquired considerable significance from an emission standpoint. The mixing pattern in this pre-ignition phase is evaluated to gain a better understanding of the processes to follow during the combustion period. In the present work, an attempt is made to bring the oxygen sensitivity into bum rate calculations via the premixed fraction, using Watson's correlation (Watson et al. 1980). The fraction of the injected fuel that burns in premixed phase p at a given speed, if related to the trapped fuel-air equivalence ratio $trapped and ignition delay period 'id in ms, is:

(9)

where k, kl , and k2 are engine-specific empirical constants. The typical recommended values of these constants are 0.9, 0.35, and 0.4, respectively, as referred to in Heywood (1988). At varying oxygen-enrichment levels, the stoichiometric oxygen-to-fuel ratio will be

(10) (mass of oxygen) (stoichiometric coefficient)

(mass offuel) OFRstoi=

4

The actual oxygen-to-fuel ratio and Tid are estimated from the multizone model developed in this work and the ignition delay equation (Equation 7), respectively. This enables a qualitative inference to be drawn from oxygen-enrichment considerations. The values of premixed burnt fraction are estimated for various levels of oxygen-enrichment, using Equation 9. These values are only representative for the study of the oxygen-enrichment effects.

Burn Rate It is well recognized that the oxygen-enrichment of standard

intake air alters the combustion characteristics of fuel-air mixtures. This change in combustion process can be viewed in two ways: (i) based on equilibrium considerations, where the equivalence ratio is the controlling parameter; and (ii) based on kinetic considerations, where both the equivalence ratio and the temperature are controlling parameters. In an equilibrium approach, the infinite reaction rate is implicit, and hence, the effect of reaction kinetics through temperature is marginal. #en oxygen-enriched air mixes with fuel, the rates of reaction increase. This effect can be better explained through the understanding of chemical reaction rates.

Assanis et al. (1993) have attempted to adopt the Double- Wiebe and Watson's burning rate calculation procedure for an oxygen-enriched air-fuel combustion system. However, they found both of these classical correlations unsatisfactory in the diffusion-burning phase. Also, in the tail region of combustion, the gradual drop in energy release rate could not be obtained. They also observed that the second function in Watson's correlation is not flexible enough to fit an oxygen-enriched combustion profile. They proposed a novel correlation with an additional degree of freedom in the original Wiebe function, in the form of a free exponent to predict the combustion tail properly. The applicability of the bum rate model suggested by Assanis et al. (1993) and also the correlations suggested by Miyamoto et al. (1985) and Huang et al. (1994) are tested for oxygen-enriched combustion in this study.

RESULTS AND DISCUSSION The typical results in the present work are obtained and

analyzed from the viewpoint of understanding of oxygen- enrichment effects on various engine processes. These results are discussed below.

Effect on Thermodynamic and Transport Properties It has been reported (Karim et al. 1965) that oxygen-

enrichment has no effect on thermodynamic and transport properties. However, a fresh estimation was undertaken to corroborate these findings. When oxygen content in the air increases from 2 1 to 3 1 % by volume, the density of air increases by 1.4% because of an increase in the RMM. Although the change in density is marginal due to variation in the RMM, this effect is of significance. where the higher molecular mass of reformed air will provide higher density at a given temperature prior to combustion. Because of the effect of an increase in the in-cylinder temperature on combustion in the oxygen-enriched air-fuel system, a decrease in density is expected, and this has

been reported (Maxwell et al. 1993). Hence, the temperature effect is a more dominant factor than the RMM for density variations with oxygen-enriched intake air. The effects of oxygen-enrichment on specific heat capacity and specific heat ratio are negligible (not shown), corroborating the results reported earlier (Karim et ai. 1965).

The variations in thermal conductivity with oxygen- enrichment were marginal, as shown in Figure 1. The slight increase in thermal conductivity with oxygen enrichment at the corresponding flame temperature explains the higher heat transfer losses, Perhaps this effect may account for a reduction in brake thermal efficiency, as reported by Wartinbee (1971) and Quader (1978), in an SI engine. On the other hand, the changes in both the dynamic and kinematic viscosities with oxygen-enrichment are negligible (not shown).

90, I

0 1000 2000 3ooo Temperalure (K)

Fig. 1 Variation of thermal conductivity with oxygen-enrichment

Effect on Flame Temperature and Equilibrium Composition The computed values of adiabatic flame temperature with

equivalence ratio at various oxygen concentrations are shown in Figure 2. It is observed that the maximum flame temperature occurs at different equivalence ratio values greater than stoichiometric when the oxygen concentration in air is increased. The maximum flame temperature is higher with higher oxygen concentration in the air, and so is the equivalence ratio for these maximum flame temperatures, as shown in Figure 2.

I 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Fuel-Air Equivalence Ratio

Fig. 2 Variation of adiabatic flame temperature with fuel-air equivalence ratio at various levels of oxygen-enrichment

Of the twelve species from equilibrium composition calculations, six species (vit., OH, N, N2, 0, 02, and NO) are of importance in NO and soot kinetics. The variations in concentration of these species with equivalence ratio at various oxygen-enrichment levels are shown in Figure 3. It is observed that an increase in oxygen content of air has increased NO, N, 0, and OH concentrations. The concentration of 0 2 is relatively constant, while N2 concentration has decreased with oxygen- enrichment, as expected. The fact that the 0 2 concentration is unaffected despite the increase in the oxygen-enrichment negates

5

1.25E-03 I I

5.00E-04

f 2.50E-04

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Fuel-Air Equivalence Ratio

;:::E ; - 3.OE-08 2.5E-08

2 8 2.0E-08 e = 1.5E-08

1.OE-08 0 5.OE-09

O.OE+OO

.- L

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

F W i Equivalence Ratio

0.4 I I

0.0 ' I 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Fueh4ir Equivalence Ratio

l.OE.04 I #

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

F u e W Equivalence Ratio

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 FueCAir Equivalence Ratio

V 4.0E-05 1

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 FueWir Equivalence Ratio - 21% oxvpen --- n%oxypen ...... 25% OxKlan

, -._. 27% Oxygen -..- 2% Oxygen

Fig. 3 Equilibrium compositions with fuel-air equivalence ratio at various oxygen-enrichment levels

the usual claim that oxygen-enrichment is akin to lean-mixture operation. Had this claim been true, the excess 02 concentration should have shown an increase in the products. In diesel combustion, the analogy of oxygen-enrichment with lean-burn conditions becomes obscure due to the heterogeneous nature of the charge. Hence, the oxygen-enrichment effect should be viewed as an effect related to the changes in nitrogen-to-oxygen ratio in air and oxygen-to-fuel ratio in the mixture, rather than as an equivalent lean air-fuel ratio. The increased OH concentration with oxygen-enrichment suggests that the HC emissions would be reduced, which has been generally reported in the literature. The peak NO concentration occurs at leaner mixtures with oxygen-enrichment than with standard air. This can be attributed to higher flame temperatures at higher oxygen concentration levels and to higher N or 0 concentrations at mixtures leaner than standard air. These equilibrium composition calculations form the basis of NO and soot oxidation calculations in the emission models. The need for an approach based on chemical kinetics is clearly felt when the oxygen-enrichment concept is examined through equilibrium considerations.

Effect on Phase Change (Fuel-Evaporation) Rate The rate at which fuel changes from the liquid to the vapor

phase, prior to combustion, is important from the mixture preparation standpoint. The effect of oxygen-enrichment on phase change rate is estimated from this standpoint; the results are shown in Figure 4. The change in fuel mass fraction (see Figure 4) at the fuel-oxidant interface has no significant effect on the mass transfer number, but changing the enthalpy transfer number influences the transfer rate (point of intersection of Bm and Bh curves). The higher the transfer number value, the higher will be the mass transfer driving potential and hence the rate. A similar trend is observed for the variation with the ratio of fuel- air interface temperature (Tw) and the critical temperature (Tcr) of the fuel (Figure 4).

12 , I

0.0 0.2 0.4 0.6 0.8 1 .o Fuel Mass Fradion

12 , I

0.6 0.7 0.8 0.9 1.0

Tenperahxe Ratio WL) -21%- _-- p%cbr/an ...... 25%- - . - - m w _..- 29%-

Fig. 4 Variation of transfer number with fuel mass fraction and temperature ratio

6

Effect on Ignition Delav Period A reduction in ignition delay with oxygen-enrichment has

been a universal experimental observation made by several investigators (Ghojel et al. 1983; Watson et al. 1990; Sekar et al. 1990; Reader et al. 1995). There is, however, no reported direct attempt to propose an ignition delay correlation incorporating the oxygen-enrichment effect. An attempt, therefore, has been made in this work to correlate the available ignition delay data. The proposed ignition delay correlation, given in Equation 8, has been used to compute the ignition delay period for varying levels of oxygen-enrichment. No attempt has been made to adjust the value of empirical constant ‘a’ appearing in the correlation to match the predicted and experimental results. Experimental data were obtained from the test results on a single-cylinder Caterpillar engine (Sekar et al. 1991). The results are presented in Figure 5. The predicted ignition delay, using the proposed correlation, shows a qualitative agreement with the experimental data but consistently underpredicts compared with experimental values at both engine load conditions. Efforts to universalize the correlation have been only partially successful and could not be pursued due to the lack of adequate published data. For that purpose, a large data set is essential; such a task may be attempted later. The qualitative trends of oxygen-enrichment effects on ignition delay with Hiroyasu’s correlation and the proposed correlation are observed to be similar. However, the quantitative data using Hiroyasu’s correlation considerably overpredicted, compared with both experimental and proposed correlation ignition delays. The changes in ‘FOR’ due to oxygen-enrichment and the variation of characteristic pressure and temperature (at the instant of injection) incorporated in the proposed correlation are attributable to the observed quantitative differences when compared to Hiroyasu’s original correlation.

2.0 I - ,20

.- 5 1.0 ti

0.8

0.6 UI

z - f 0.4 U.

2 0.2

6 0.0 m I .- -

0 5 10 15 20 25 30

Crank Angle CCA)

Fig. 6 Variation of ignitable fuel mass fraction with injection duration for various levels of oxygen-enrichment

.r,c--.+.-.-.---. * - - - .. .>..+./ .:-: :- ‘ - -

Ignition delay line

21WOxygen , ...... 23% Oxygen -..- 25% 27%Oxygen Oxygen 1 --- 29%Oxygen ,

ignition delay period presented in Table 2 appears to be long, as computed using Hiroyasu’s correlation (Equation 7) instead of the proposed correlation (Equation 8). Because it is difficult to know ‘FOR’ at the point of ignition, the delay period itself needs to be estimated. For quantitative comparisons, more experimental data are being sought to predict the ignition delay period and the corresponding premixed burnt fraction for various oxygen-enrichment levels. The ignitable fuel mass fraction is estimated by using a model that predicts the charge heterogeneity (spatial distribution). At a given oxygen concentration level, the variation of the premixed burnt fraction with ignitable fuel mass fraction prepared during the ignition delay period is shown in Figure 7. This figure indicates a very useful relationship between premixed burnt mass fraction and the ignitable fuel mass fraction prepared during the delay period.

It is also observed (see Figure 8) that increasing oxygen- enrichment reduces the premixed burnt fraction. This suggests that there will be a smoother pressure rise with oxygen- enrichment. This expectation is in conformity with reported observations in several experimental studies about reduced peak pressures with oxygen-enrichment (Watson et al. 1990; Sekar et

0.0 1 M 24 28 32 36

Fig. 5 Variation of ignition delay with oxygen-enrichment

Effect on Fuel-Air Mixture Quality The variation of ignitable fuel mass fraction prepared during

the ignition delay period for different levels of oxygen- enrichment obtained from the multizone, spray-mixing model is shown in Figure 6. The ignitable fuel mass fractions have been estimated from the start of injection to the instant of ignition. Hence, if the instants of ignition for varying levels of oxygen- enrichment are plotted. as in Figure 6, a delay line is obtained, which represents the instant of ignition at each oxygen level.

The estimated values of the premixed burnt fraction for various oxygen-enrichment levels are summarized in Table 2, along with other useful parameters referred to in this work. The

0.66 ,

/

0.75 0. Bo 0.85 0.90 0.95 1.00 Ignitable Fuel Mass Fraction

Fig. 7 Variation of premixed burnt fraction with ignitable mass fraction

0.66 r c I .? 0.60

5 0.48 0.54

m 3 0.42 .I

0.36

0.30 a

19 21 23 25 27 29 31 33 Oxygen Concentration (percent volume)

Fig. 8 Variation of premixed burnt fraction at various levels of oxygen-enrichment

7

Table 2. Summary of results predicting the effects of oxygen-enrichment Oxygen Stoichiometric Ignition delay Trapped Trapped Premixed burnt Ignitable fuel (volume oxygen-to-fuel (degree crank oxygen-to-fuel equivalence fraction 'p' mass fraction

23 25 27 29 31

3.814 4.146 4.477 4.808 5.140

19 17 15 13 12

10.504 9.962 9.93 8 8.775 8.455

0.363 0.416 0.450 0.547 0.607

0.519 0.473 0.430 0.354 0.336

0.920 0.907 0.875 0.834 0.777

al. 1990: Assanis et al. 1993). The ability to correlate model parameters with this feature elucidates the usefulness of the present work. The stoichiometric contour of the oxygen-to-fuel ratio can be considered as a typical representation of the mixing pattern in the spray. Figure 9 shows a plot of stoichiometric contours of the oxygen-to-fuel ratio at various oxygen- enrichment levels. This figure suggests that the movement of the burning zone is characterized by the stoichiometric contour of the oxygen-to-fuel ratio. It can be seen that the burning zone will be closer to the edge of the spray than to its center with increased oxygen-enrichment. This variation is significant because it alters the local burning characteristics in various spray zones around the stoichiometric contour within the flammability limits. This may possibly reveal a suitable relationship between combustion and emission characteristics. This relationship must be further investigated and established as charge heterogeneity is duly accounted for.

0 5 10 Normalied Radial Position

15

Fig. 9 Spatial variation of stoichiometric oxygen-to-fuel ratio

Effect on Burninp Rate The correlation suggested by Assanis et al. (1993) required the

changing of empirical constants with respect to engine geometry, operating conditions, and oxygen-enrichment level to obtain a satisfactory bum rate profile when compared with experimental data. This suggests the engine-specific nature of the correlation. Attempts to use the correlation proposed by Miyamoto et al. (1985) resulted in overprediction of the premixed combustion phase. The duration of the premixed phase is not constant; it varies with operating conditions and engine geometry. Also, the instantaneous rate of burning at different crank angles deviates from experimental data. When the correlation proposed by Huang et al. (1994) is used to evaluate the bum rate profile, it is found that it consistently overpredicts the premixed burning rate and underpredicts the diffusion-burning rate. This implies that

the generality claimed for the equations of Huang et al. (1994) with regard to Wiebe's parameters does not exist. The approaches of Assanis et al. (1993), Miyamoto et al. (1985), and Huang et al. (1994) are too empirical to be extended to any generalized modeling of oxygen-enrichment. Since each burning rate correlation requires calibration with specific experimental data, the burn rates from these correlations were difficult to compare. It seems that a multidimensional or a multizone, phenomenological spray combustion approach may be more appropriate to describe the burning rate when oxygen-enriched combustion air is used. From the standpoint of computation and concept validation, the usefulness of a multizone, phenomenological spray mixing approach has been established in the present work for the premixed phase of combustion. The spray combustion model can be extended to prediction of complete combustion and emission characteristics.

CONCLUSIONS In the present work, a process-based, time-marching multizone

model for a quiescent DI diesel engine has been developed, and a correlation for ignition delay and a qualitative relationship of the premixed burnt fraction with oxygen-enrichment have been successfully achieved. In addition, fundamental insight has been gained concerning the effects of oxygen-enrichment on engine combustion. In summary, the effects of oxygen-enrichment are:

1.

2.

3.

4.

5.

6.

Negligible effects on specific heat capacity, the ratio of specific heats, and dynamic and kinematic viscosity. The increase in charge density due to variation in RMM is very slight (< 2%). Marginal increase in thermal conductivity with increasing oxygen-enrichment for any given flame temperature. Higher maximum adiabatic flame temperatures with higher oxygen concentrations in air. The fiel-air equivalence ratios are also higher for the corresponding maximum flame temperatures. Insignificant effect on mass transfer number, but an increase in enthalpy transfer number. Reduction in ignition delay. The ignition delay correlation proposed in terms of fuel-to-oxygen ratio is:

Considerable reduction in premixed burnt fraction.

8

c .

ACKNOWLEDGMENTS Mehta, P.S., D. Lahiri, and T. Bhaskar, 1996. Evaluation of The US. Department of Energy, Office of Energy Research, Role of Premixed Combustion on Diesel Engine Combustion and

supported this work under contract W-3 1-109-Eng-38. The Emission Characteristics, SAE Paper No. 962482. contributions of A.V.R.C. Prasad and T. Bhasker are gratefully Miyamoto, N., T. Chikasha, T. Murayama, and R. Sawyer, acknowledged. 1985, Description and Analysis of Diesel Engine Rate of

Combustion and Performance Using Wiebe 's Functions, SAE Paper No. 850107.

REFERENCES Olikara, C., and G.L. Barman, 1975, A Computer Program for Abramovich, G.N., 1963, The Theory of Turbulent Jets, MIT Calculating Properties of Equilibrium Combustion Products with

Press, Cambridge, Mass. Some Application to IC Engines, SAE Paper No. 750468. Andree, A., and S.J. Pachernegg, 1969, Ignition Conditions in Quader, A.A., 1978, Exhaust Emissions and Performance of a

Diesel Engines, S A E Paper No. 690253. Spark-Ignition Engine Using Oxygen Enriched Intake Air, Assanis, D.N., E. Karvounis, R.R. Sekar, and W.W. Marr, Combustion Science and Technology, Vol. 19, pp. 81-86.

1993, Heat Release Analysis of Oxygen-Enriched Diesel Reader, G.T., C.E. Bowen, P. Maier, I.J. Potter, and R.W. Combustion, ASME Journal of Engineering for Gas Turbines and Gustafson, 1995, The Effects of Oxygeflitrogen Ratios on the Power, Vol. 115, pp. 761-768. Performance and Emissions of an IDI Diesel Engine, Third

Dent, J.C., 1971, A Basis for the Comparison of Various Asian-Pacific International Symposium on Combustion and Experimental Methods for Studying Spray Penetration, S A E Energy Utilization. PaperNo. 710571,1971. Ricou, F.P., and D.B. Spalding, 1961, Measurements of

Fujimoto, H., 1980, Illumination Delay in Diesel Spray, JSAE Entrainment by Axisymmetrical Turbulent Jets, J. Fluid Paper No. 800-13. Mechanics, Vol. 11.

Ghojel, J., J.C. Hilliard, and J.A. Levendis, 1983, Effect of Rosner, D.E., and W.S. Chang, 1973, Transient Evaporation Oxygen-Enrichment on the performance of IDI Diesel Engines, and Combustion of a Fuel Droplet Near its Critical Temperature, S A E Paper No. 830245. Combustion Science and Technology, Vol. 7.

Heywood, J.B., 1988, Internal Combustion Engine Sekar, R.R., W.W. Marr, R.L. Cole, T.J. Marciniak, and J.E. Fundamentals, McGraw-Hill Publications, New York, N.Y. Schaus, 1990, Diesel Engine Experiments with Oxygen

Hiroyasu, H., 1985, Diesel Engine Combustion and its Enrichment, Water Addition, and Lower-Grade Fuel, ASME Modeling, Proceedings of the COMODIA-85, pp. 53-85. Energy Sources Technology Conference and Exhibition, Paper

Release Patterns of High Sped DI Diesel Engines, S A E Paper Sekar, R.R., W.W. Marr, R.L. Cole, and T.J. Marciniak, 1991, No. 941694. Experimental Evaluation of Oxygen-Enriched Air and Emulsified

Iida, N., Y. Suzuki, G.T. Sato, and T. Sawada, 1986, Effects of Fuels in a Single-Cylinder Diesel Engine, Technical Report, Intake w e n Concentration on the Characteristics of ANLESDKM-30, Vols. 1 and 2, Argonne National Laboratory. Particulate Emissions @om D.I. Diesel Engine, S A E Paper No. Shahed, S.M., W.S. Chiu, and W.T. Lyn, 1975, A 86 1233. Mathematical Model for Diesel Combustion, Proceedings of the

Performance of a Compression-Ignition Engine in Spalding, D.B., 1963, Convective Mass Transfer, Edward Unconventional Atmospheres, The Engineer, No. 5695, N 219, Arnold (Publishers) Limited, London (U.K). March 26, pp. 551-556. Sparrow, E.M., and J.L. Gregg, 1958, The Variable Fluid

Karim, G.A., and G. Ward, 1968, The Examination of the Properv Problem in Free Convection, ASME Transactions, Vol.

Changing the Partial Pressure of Oxygen in the Intake Charge, Virk, S.K., K. Uygur, and C.R. Bartels, 1993, E#ects of SAE Paper No. 680767. Oxygen-Enriched Air in Diesel Engine Exhaust Emissions and

Kothandaraman, C.P., and S. Subramanyam, 1989, Heat and Engine Performance, S A E Paper No. 93 1004. Mass Transfir Data Book, Fourth Edition, Wiley Eastern Wartinbee, W.J. Jr., 1971, Emissions Study of Ovgen- Limited, New Delhi (India). Enriched Air, SAE Paper No. 710606.

Maxwell, T.T., V. Setty, J.C. Jones, and R. Narayan, 1993, Watson, H.C., E.E. Milkins, and G.R. Rigby, 1990, A New The Eflect of Oxygen-Enriched Air on the Performance and Look at Oxygen Enrichment 1) The Diesel Engine, SAE Paper Emissions of Internal Combustion Engines, S A E Paper No. No. 900344. 932804. Watson, N., A.D. Pilley, and M.A. Marzouk, 1980,

Mehta, P.S., J.C. Dent, and A.K. Gupta, 1985, Computational Combustion Correlation for Diesel Engine Simulation, SAE Schemes for Atomization and Evaporation of Diesel Fuel Spray, Paper No. 800029.

Huang, Y., L. Zhang, and W. Lu, 1994, Optimum Heat NO. 90-ICE-1.

Karim, G.A., D.E. Gee, and R.T.C. Satterford, 1965, Institute of Mechanical Engineers.

Combustion Processes in a Compression-Ignition Engine by 80, pp. 879-886.

IB/3/1-10, ICLASS-85.

9